AI usage is rocketing, while Google’s share of the global search market recently dipped below 90% for the first time since 2015. So does this mean it's the end of search as we know it?
Short answer: No, not really – at least, not yet.
We love a deep dive here at Adido, so we grabbed our tools and dug into the data across 31 travel websites with a total of just over 4 million visits so far this year. Out of those visits, just 4,769 could be clearly traced back to AI sources. That's 0.12% - or to put it a different way, about 12 visits in every 10,000. You're more likely to get traffic from DuckDuckGo (yes, that's still alive and quacking), which in some cases drove 1% or more.
And when it comes to which LLM is reigning supreme, there's really only one player in town. ChatGPT accounts for a whopping 86.4% of all identifiable LLM referrals. The rest? Barely a blip.
Now, we know not all AI-driven sessions show up neatly as referrals so it's fair to say we're not getting the full picture. But even with that caveat, no one can argue that the AI gold rush hasn't quite hit the travel sector just yet.
There are a few exceptions. Some US-focused sites in our batch did see ever-so-slightly higher percentages – over 0.5% in a couple of cases. But again, that's still tiny.
Comparing apples and oranges?
Whilst we now know that visits to travel websites from LLMs like ChatGPT are currently very low, it’s perhaps a little unfair to measure their impact in the traditional sense. The core purpose of LLMs and AI bots is to provide answers, not drive clicks. So expecting a surge in traffic from the millions of questions asked each day may not be a realistic comparison. However, to put things into broader perspective, if we worked backwards and tried to estimate how often a brand name appeared in responses, we might get a better sense of whether the tide is turning (or not).
It would seem that there are no or very few studies showing how many LLM searches result in a click. In my quest to find out, I asked Claude.ai ‘how many LLM searches result in a click to a website?’, hoping that the LLM would know about its own behaviour. And indeed, I was served up with some interesting albeit pithy data saying ‘anything from 10% to 70% of searches result in a click to a website’. Hmmm. Not very helpful. It did caveat its response by saying that the click rate would vary a lot depending on if it was an easy or hard question which does make sense. When asked where the source of this data was to find out more… well, you can see for yourself.
So without using made up data or any other published information, we have to perhaps look to the current stats about Google click data which show that roughly 60% or so of searches don’t end up in a click to a website (although that study is nearly a year old and way before AIOs really kicked in).
A study from Bain last year showed some similar data with 80% of searches relying on AI summaries 40% of the time (yeah, not the easiest stat to get your head around I know…). If we average it out, let’s say roughly 50% get an answer and don’t click. By that logic, perhaps our 0.12% figure should be doubled to account for ‘no-click’ answers that still feature travel brands. We’ll never really know until one of the LLMs publishes some real data about how many brands are returned in the results that are served up (a ChatGPT Console perhaps?) and perhaps they won’t do this for a good while in case the numbers spook the market/medias/competitors.
But you know what they say - the early bird catches the worm and all that. So for those marketers who are keen to be matchfit when (or if) AI search properly takes off, there are steps you can take to increase your chances of being found in the LLMs. Let’s take a closer look at some of these strategies below.
How to increase your chance of appearing in AI search results
1. Create concise content that answers people’s questions
AI tools are designed to serve up quick, useful answers. So if someone types ‘best family holidays in Crete’, your page should give them what they came for, keeping waffle to an absolute minimum. Add practical tips, FAQs and even tables or bullet points to help AI pick out the useful stuff.
2. Refresh your pages regularly
Just like search engines, AI doesn’t like outdated content. If your blog on ‘Summer 2023 travel tips’ is collecting dust online, update it for this year with current prices, new flight routes or any recent visa changes. This tells AI (and Google) that your content’s still fresh as a daisy.
3. Optimise your Google Business Profile
This is especially important for local travel agents. Make sure your Google Business Profile (formally Google My Business) listing is up to date and accurate – think address, contact details, opening hours, services and photos. AI tools use this data when building quick answer cards or local search results. Don’t forget to respond to reviews and answer any FAQs too.
4. Target longer, more specific search terms
You’re more likely to appear in AI-generated answers if you go after detailed queries such as ‘cheap wine tasting tours for couples in Sicily’ rather than broad, highly competitive ones like ‘wine tasting holidays’. Home in on your niche and your brand’s USP.
5. Focus on E-E-A-T
Google frequently emphasises the importance of E-E-A-T (Experience, Expertise, Authoritativeness, Trustworthiness) and prioritises content on its platform that demonstrate these credentials. The same is applicable for AI referrals. LLMs love user-generated content, especially reviews. Encourage customers to leave feedback – good and bad – on Google, Trustpilot or Tripadvisor. Reply to any questions and comments too – these interactions show up in AI crawls and can boost your chances of being quoted.
Final thoughts
All the talk right now is how the world of search is changing. Usage of tools like ChatGPT continues to grow rapidly, reaching 800m weekly visits as of April 2025 - double what it was only a few months ago. For now though, AI search referrals aren’t a game changer - but they are interesting to watch. If nothing else, it’s a good barometer for how fast - or not – consumer habits are changing.
We’ll keep tracking the data and shout up when the numbers get exciting. For now, though, it’s business as usual.
